Latent Feature Representation Research Articles

TRust Your GENerator (TRYGEN): Enhancing Out-of-Model Scope Detection

Recent research has drawn attention to the ambiguity surrounding the definition and learnability of Out-of-Distribution recognition. Although the original problem remains unsolved, the term “Out-of-Model Scope” detection offers a clearer perspective. The ability to detect Out-of-Model Scope inputs is particularly beneficial in safety-critical applications such as autonomous driving or medicine. By detecting Out-of-Model Scope situations, the system’s robustness is enhanced and it is prevented from operating in unknown and unsafe scenarios. In this paper, we propose a novel approach for Out-of-Model Scope detection that integrates three sources of information: (1) the original input, (2) its latent feature representation extracted by an encoder, and (3) a synthesized version of the input generated from its latent representation. We demonstrate the effectiveness of combining original and synthetically generated inputs to defend against adversarial attacks in the computer vision domain. Our method, TRust Your GENerator (TRYGEN), achieves results comparable to those of other state-of-the-art methods and allows any encoder to be integrated into our pipeline in a plug-and-train fashion. Through our experiments, we evaluate which combinations of the encoder’s features are most effective for discovering Out-of-Model Scope samples and highlight the importance of a compact feature space for training the generator.

Structural topology optimization based on deep learning

In the mechanical design of structures, traditional topology optimization methods involve numerous finite element iterative analyses, leading to a significant expenditure of computational resources. Therefore, the improved multi-scale gradient generative adversarial networks topology optimization technique is proposed. The topology optimization condition parameters are compressed into a low-dimensional latent space feature representation using the encoder, allowing the model to better extract features from these parameters. To speed up model training, the generator and discriminator networks use lightweight residual convolutional blocks. The hybrid attention mechanism extracts prominent region features from the topology optimization structure map. The model training process is guided by a multi-dimensional fusion loss function to enhance the quality of generated model samples. Finally, transferring the parameters of the low-resolution topology optimization model to the high-resolution model enables complete training on a limited amount of high-resolution topology optimization datasets. The experimental data on the low- and high-resolution topology optimization datasets demonstrate that, when compared to alternative methods, this method produces better-quality topology optimization structure maps. Additionally, it can generate high-resolution topology optimization structure maps in minimal time, enabling real-time topology optimization.

Deep coupled registration and segmentation of multimodal whole-brain images.

Recent brain mapping efforts are producing large-scale whole-brain images using different imaging modalities. Accurate alignment and delineation of anatomical structures in these images are essential for numerous studies. These requirements are typically modeled as two distinct tasks: Registration and segmentation. However, prevailing methods, fail to fully explore and utilize the inherent correlation and complementarity between the two tasks. Furthermore, variations in brain anatomy, brightness and texture pose another formidable challenge in designing multi-modal similarity metrics. A high-throughput approach capable of overcomingthe bottleneck of multi-modal similarity metric design, while effective leveraging the highly correlated and complementary nature of two tasks is highly desirable. We introduce a deep learning framework for joint registration and segmentation of multi-modal brain images. Under this framework, registration and segmentation tasks are deeply coupled and collaborated at two hierarchical layers. In the inner layer, we establish a strong feature-level coupling between the two tasks by learning a unified common latent feature representation. In the outer layer, we introduce a mutually supervised dual-branch network to decouple latent features and facilitate task-level collaboration between registration and segmentation. Since the latent features we designed are also modality-independent, the bottleneck of designing multi-modal similarity metric is essentially addressed. Another merit offered by this framework is the interpretability of latent features, which allows intuitive manipulation of feature learning, thereby further enhancing network training efficiency and the performance of both tasks. Extensive experiments conducted on both multi-modal and mono-modal datasets of mouse and human brains demonstrate the superiority of our method. The code is available at https://github.com/tingtingup/DCRS. Supplementary data are available at Bioinformaticsonline.

Joint self-supervised and supervised contrastive learning for multimodal MRI data: Towards predicting abnormal neurodevelopment

The integration of different imaging modalities, such as structural, diffusion tensor, and functional magnetic resonance imaging, with deep learning models has yielded promising outcomes in discerning phenotypic characteristics and enhancing disease diagnosis. The development of such a technique hinges on the efficient fusion of heterogeneous multimodal features, which initially reside within distinct representation spaces. Naively fusing the multimodal features does not adequately capture the complementary information and could even produce redundancy. In this work, we present a novel joint self-supervised and supervised contrastive learning method to learn the robust latent feature representation from multimodal MRI data, allowing the projection of heterogeneous features into a shared common space, and thereby amalgamating both complementary and analogous information across various modalities and among similar subjects. We performed a comparative analysis between our proposed method and alternative deep multimodal learning approaches. Through extensive experiments on two independent datasets, the results demonstrated that our method is significantly superior to several other deep multimodal learning methods in predicting abnormal neurodevelopment. Our method has the capability to facilitate computer-aided diagnosis within clinical practice, harnessing the power of multimodal data. The source code of the proposed model is publicly accessible on GitHub: https://github.com/leonzyzy/Contrastive-Network.

GIAE-DTI: Predicting Drug-Target Interactions Based on Heterogeneous Network and GIN-based Graph Autoencoder.

Accurate prediction of drug-target interactions (DTIs) is essential for advancing drug discovery and repurposing. However, the sparsity of DTI data limits the effectiveness of existing computational methods, which primarily focus on sparse DTI networks and have poor performance in aggregating information from neighboring nodes and representing isolated nodes within the network. In this study, we propose a novel deep learning framework, named GIAE-DTI, which considers cross-modal similarity of drugs and targets and constructs a heterogeneous network for DTI prediction. Firstly, the model calculates the cross-modal similarity of drugs and proteins from the relationships among drugs, proteins, diseases, and side effects, and performs similarity integration by taking the average. Then, a drug-target heterogeneous network is constructed, including drug-drug interactions, protein-protein interactions, and drug-target interactions processed by weighted K nearest known neighbors. In the heterogeneous network, a graph autoencoder based on a graph isomorphism network is employed for feature extraction, while a dual decoder is utilized to achieve better self-supervised learning, resulting in latent feature representations for drugs and targets. Finally, a deep neural network is employed to predict DTIs. The experimental results indicate that on the benchmark dataset, GIAE-DTI achieves AUC and AUPR scores of 0.9533 and 0.9619, respectively, in DTI prediction, outperforming the current state-of-the-art methods. Additionally, case studies on four 5-hydroxytryptamine receptor-related targets and five drugs related to mental diseases show the great potential of the proposed method in practical applications.

An informative dual ForkNet for video anomaly detection

An autoencoder for video anomaly detection task is a type of algorithm with the primary purpose of learning an “informative” representation of the normal data that can be used for identifying the abnormal data by learning to reconstruct a set of input observations. Based on the encoding–decoding structure, we explore a novel dual ForkNet architecture that can dissociate and process the spatio-temporal representation. It is well-known in the information theory community that most autoencoders coding processes are inevitably accompanied by a certain loss of information. In this dual ForkNet, we focus on mitigating the information loss problem and propose a novel architectural recalibration approach, which we term the “Informetrics Recalibration” (IR). It can adaptively recalibrate latent feature representation by explicitly modeling the similarity between the corresponding feature maps of encoder and decoder, and retain more useful semantic information to generate greater differentiation between normal and abnormal events. Additionally, because the structure of the autoencoder itself determines the difficulty to obtain deep semantic information, we introduce a Secondary Encoder (SE) in each ForkNet, so as to recalibrate target features responses of latent feature representation. Our model is easy to be trained and robust to be applied, because it basically consists of some ResNet blocks without using complicated modules. Extensive experiments on the five publicly available benchmarks show that our model outperforms the existing state-of-the-art architectures, demonstrating our framework’s effectiveness.

Cube is a good form: Hyperspectral band selection via multi-dimensional and high-order structure preserved clustering

As an effective strategy for reducing the noisy and redundant information for hyperspectral imagery (HSI), hyperspectral band selection intends to select a subset of original hyperspectral bands, which boosts the subsequent different tasks. In this paper, we introduce a multi-dimensional high-order structure preserved clustering method for hyperspectral band selection, referred to as MHSPC briefly. By regarding original hyperspectral images as a tensor cube, we apply the tensor CP (CANDECOMP/PARAFAC) decomposition on it to exploit the multi-dimensional structural information as well as generate a low-dimensional latent feature representation. In order to capture the local geometrical structure along the spectral dimension, a graph regularizer is imposed on the new feature representation in the lower dimensional space. In addition, since the low rankness of HSIs is an important global property, we utilize a nuclear norm constraint on the latent feature representation matrix to capture the global data structure information. Different to most of previous clustering based hyperspectral band selection methods which vectorize each band as a vector without considering the 2-D spatial information, the proposed MHSPC can effectively capture the spatial structure as well as the spectral correlation of original hyperspectral cube in both local and global perspectives. An efficient alternatively updating algorithm with theoretical convergence guarantee is designed to solve the resultant optimization problem, and extensive experimental results on four benchmark datasets validate the effectiveness of the proposed MHSPC over other state-of-the-arts.

MambaNet: A Hybrid Neural Network for Predicting the NBA Playoffs

In this paper, we present MambaNet: a hybrid neural network for predicting the outcomes of Basketball games. Contrary to other studies, which focus primarily on season games, this study investigates playoff games. MambaNet is a hybrid neural network architecture that processes a time series of teams’ and players’ game statistics and generates the probability of a team winning or losing an NBA playoff match. In our approach, we utilize Feature Imitating Networks to provide latent signal-processing feature representations of game statistics to further process with convolutional, recurrent, and dense neural layers. Three experiments using five different datasets are conducted to evaluate the performance and generalizability of our architecture against a wide range of previous studies. Our final method successfully predicted the AUC from 0.72 to 0.82, beating the best-performing baseline models by a considerable margin.

Detecting Logos for Indoor Environmental Perception Using Unsupervised and Few-Shot Learning

Indoor scenes are crucial components of urban spaces, with logos serving as vital information within these environments. The accurate perception of logos is essential for effectively operating mobile robots in indoor environments, which significantly contributes to many upper-level applications. With the rapid development of neural networks, numerous deep-learning-based object-detection methods have been applied to logo detection. However, most of these methods depend on large labeled datasets. Given the fast-changing nature of logos in indoor scenes, achieving reliable detection performance with either the existing large labeled datasets or a limited number of labeled logos remains challenging. In this article, we propose a method named MobileNetV2-YOLOv4-UP, which integrates unsupervised learning with few-shot learning for logo detection. We develop an autoencoder to obtain latent feature representations of logos by pre-training on a public unlabeled logo dataset. Subsequently, we construct a lightweight logo-detection network and embed the encoder weights as prior information. Training is performed on a small dataset of labeled indoor-scene logos to update the weights of the logo-detection network. Experimental results on the public logo625 dataset and our self-collected LOGO2000 dataset demonstrate that our method outperforms classic object-detection methods, achieving a mean average detection precision of 83.8%. Notably, our unsupervised pre-training strategy (UP) has proven effective, delivering a 15.4% improvement.

Indexing ECG for Integrated Health Social Networks Predicting Keywords from ECG to Access Online Information

Health Social Networks (HSN) provide rich medical knowledge bases that are scalable and sustainable, while IoT provides non-invasive, pervasive, and low-cost methods to collect patient data. However, receiving relevant information from HSN is time consuming and challenging for users, such as searching for the right relevant information using keywords and filtering. On the other hand, healthcare IoT has limited access to the vast medical knowledge bases, such as HSN, to interpret the collected data. To address these challenges, we propose Keyword-based Integrated HSN of Things (KIHoT), an approach that combines the strengths of both HSNs and IoT to overcome their limitations. In this method, data (biosignals) collected via IoT devices are converted to human readable keywords using word embedding vector features and CNN (Convolutional Neural Network) predictors. The CNN predictors are trained to predict keywords that individuals search within an HSN to extract relevant information of the given biosignals. Those keywords are encoded as word embedding for searching relevant information. KIHoT utilizes contrast learning techniques to extract latent feature representations of electrocardiogram (ECG) signals, which are then used to predict disease-related keywords. The proposed method was evaluated using 11,936 ECG signals from patients with heart disease and achieved an accuracy of 98% for disease prediction. Our results suggest that KIHoT can effectively extract relevant information from HSN portals, making it easier for researchers and clinicians to access valuable medical knowledge.

Semi-supervised segmentation of abdominal organs and liver tumor: uncertainty rectified curriculum labeling meets X-fuse

Precise liver tumors and associated organ segmentation hold immense value for surgical and radiological intervention, enabling anatomical localization for pre-operative planning and intra-operative guidance. Modern deep learning models for medical image segmentation have evolved from convolution neural networks to transformer architectures, significantly boosting global context understanding. However, accurate delineation especially of hepatic lesions remains an enduring challenge due to models’ predominant focus solely on spatial feature extraction failing to adequately characterize complex medical anatomies. Moreover, the relative paucity of expertly annotated medical imaging data restricts model exposure to diverse pathological presentations. In this paper, we present a three-phrased cascaded segmentation framework featuring an X-Fuse model that synergistically integrates spatial and frequency domain’s complementary information in dual encoders to enrich latent feature representation. To enhance model generalizability, building upon X-Fuse topology and taking advantage of additional unlabeled pathological data, our proposed integration of curriculum pseudo-labeling with Jensen–Shannon variance-based uncertainty rectification promotes optimized pseudo supervision in the context of semi-supervised learning. We further introduce a tumor-focus augmentation technique including training-free copy-paste and knowledge-based synthesis that show efficacy in simplicity, contributing to the substantial elevation of model adaptability on diverse lesional morphologies. Extensive experiments and modular evaluations on a holdout test set demonstrate that our methods significantly outperform existing state-of-the-art segmentation models in both supervised and semi-supervised settings, as measured by the Dice similarity coefficient, achieving superior delineation of bones (95.42%), liver (96.26%), and liver tumors (89.53%) with 16.41% increase comparing to V-Net on supervised-only and augmented-absent scenario. Our method marks a significant step toward the realization of more reliable and robust AI-assisted diagnostic tools for liver tumor intervention. We have made the codes publicly available [https://github.com/lyupengju/X-Fuse].

Deep matrix factorization via feature subspace transfer for recommendation system

The sparsity problem remains a significant bottleneck for recommendation systems. In recent years, deep matrix factorization has shown promising results in mitigating this issue. Furthermore, many works have improved the prediction accuracy of deep matrix factorization by incorporating the user’s and/or items’ auxiliary information. However, there are still two remaining drawbacks that need to be addressed. First, the initialization of latent feature representations has a substantial impact on the performance of deep matrix factorization, and most current models utilize a uniform approach to this initialization, constraining the model’s optimization potential. Secondly, many existing recommendation models lack versatility and efficiency in transferring auxiliary information from users or items to expand the feature space. This paper proposes a novel model to address the issues mentioned above. By using a semi-autoencoder, the pre-trained initialization of the latent feature representation is realized in this paper. Simultaneously, this model assimilates auxiliary information, like item attributes or rating matrices from diverse domains, to generate their latent feature representations. These representations are then transferred to the target task through subspace projection distance. With this, this model can utilize auxiliary information from various sources more efficiently and this model has better versatility. This is called deep matrix factorization via feature subspace transfer. Numerical experiments on several real-world data show the improvement of this method compared with state-of-the-art methods of introducing auxiliary information about items. Compared with the deep matrix factorization model, the proposed model can achieve 6.5% improvement at most in the mean absolute error and root mean square error.

Cross noise level PET denoising with continuous adversarial domain generalization

Objective. Performing positron emission tomography (PET) denoising within the image space proves effective in reducing the variance in PET images. In recent years, deep learning has demonstrated superior denoising performance, but models trained on a specific noise level typically fail to generalize well on different noise levels, due to inherent distribution shifts between inputs. The distribution shift usually results in bias in the denoised images. Our goal is to tackle such a problem using a domain generalization technique. Approach. We propose to utilize the domain generalization technique with a novel feature space continuous discriminator (CD) for adversarial training, using the fraction of events as a continuous domain label. The core idea is to enforce the extraction of noise-level invariant features. Thus minimizing the distribution divergence of latent feature representation for different continuous noise levels, and making the model general for arbitrary noise levels. We created three sets of 10%, 13%–22% (uniformly randomly selected), or 25% fractions of events from 97 18F-MK6240 tau PET studies of 60 subjects. For each set, we generated 20 noise realizations. Training, validation, and testing were implemented using 1400, 120, and 420 pairs of 3D image volumes from the same or different sets. We used 3D UNet as the baseline and implemented CD to the continuous noise level training data of 13%–22% set. Main results. The proposed CD improves the denoising performance of our model trained in a 13%–22% fraction set for testing in both 10% and 25% fraction sets, measured by bias and standard deviation using full-count images as references. In addition, our CD method can improve the SSIM and PSNR consistently for Alzheimer-related regions and the whole brain. Significance. To our knowledge, this is the first attempt to alleviate the performance degradation in cross-noise level denoising from the perspective of domain generalization. Our study is also a pioneer work of continuous domain generalization to utilize continuously changing source domains.

Deep neural network for damage detection in Infante Dom Henrique bridge using multi-sensor data

This paper proposes a data-driven approach to detect damage using monitoring data from the Infante Dom Henrique bridge in Porto. The main contribution of this work lies in exploiting the combination of raw measurements from local (inclinations and stresses) and global (eigenfrequencies) variables in a full-scale structural health monitoring application. We exhaustively analyze and compare the advantages and drawbacks of employing each variable type and explore the potential of combining them. An autoencoder-based deep neural network is employed to properly reconstruct measurements under healthy conditions of the structure, which are influenced by environmental and operational variability. The damage-sensitive feature for outlier detection is the reconstruction error that measures the discrepancy between current and estimated measurements. Three autoencoder architectures are designed according to the input: local variables, global variables, and their combination. To test the performance of the methodology in detecting the presence of damage, we employ a finite element model to calculate the relative change in the structural response induced by damage at four locations. These relative variations between the healthy and damaged responses are employed to affect the experimental testing data, thus producing realistic time-domain damaged measurements. We analyze the receiver operating characteristic curves and investigate the latent feature representation of the data provided by the autoencoder in the presence of damage. Results reveal the existence of synergies between the different variable types, producing almost perfect classifiers throughout the performed tests when combining the two available data sources. When damage occurs far from the instrumented sections, the area under the curve in the combined approach increases [Formula: see text] compared to using local variables only. The classificatoin metrics also demonstrate the enhancement of combining both sources of data in the damage detection task, reaching close to [Formula: see text] precision values for the four considered test damage scenarios. Finally, we also investigate the capability of local variables to localize the damage, demonstrating the potential of including these variables in the damage detection task.

Switching Self-Attention Text Classification Model with Innovative Reverse Positional Encoding for Right-to-Left Languages: A Focus on Arabic Dialects

Transformer models have emerged as frontrunners in the field of natural language processing, primarily due to their adept use of self-attention mechanisms to grasp the semantic linkages between words in sequences. Despite their strengths, these models often face challenges in single-task learning scenarios, particularly when it comes to delivering top-notch performance and crafting strong latent feature representations. This challenge is more pronounced in the context of smaller datasets and is particularly acute for under-resourced languages such as Arabic. In light of these challenges, this study introduces a novel methodology for text classification of Arabic texts. This method harnesses the newly developed Reverse Positional Encoding (RPE) technique. It adopts an inductive-transfer learning (ITL) framework combined with a switching self-attention shared encoder, thereby increasing the model’s adaptability and improving its sentence representation accuracy. The integration of Mixture of Experts (MoE) and RPE techniques empowers the model to process longer sequences more effectively. This enhancement is notably beneficial for Arabic text classification, adeptly supporting both the intricate five-point and the simpler ternary classification tasks. The empirical evidence points to its outstanding performance, achieving accuracy rates of 87.20% for the HARD dataset, 72.17% for the BRAD dataset, and 86.89% for the LABR dataset, as evidenced by the assessments conducted on these datasets.

Location-Aware Deep Interaction Forest for Web Service QoS Prediction

With the rapid development of the web service market, the number of web services shows explosive growth. QoS is an important factor in the recommendation scene; how to accurately recommend a high-quality service for users among the massive number of web services has become a tough problem. Previous methods usually acquired feature interaction information by network structures like DNN to improve the QoS prediction accuracy, but this generates unnecessary computations. Aiming at addressing the above problem, inspired by the multigrained scanning mechanism in a deep forest, we propose a location-aware deep interaction forest approach for web service QoS prediction (LDIF). This approach offers the following innovations: The model fuses the location similarity of users and services as a latent feature representation of them. In addition, we designed a scanning interaction structure (SIS), which obtains multiple local feature combinations from the interaction between user and service features, uses interactive computing to extract feature interaction information, and concatenates the feature interaction information with original features, which aims to enhance the dimension of the features. Equipped with these, we compose a layer-by-layer cascade by using SIS to fuse low- and high-order feature interaction information, and the early-stop mechanism controls the cascade depth to avoid unnecessary computation. The experiments demonstrate that our model outperforms eight other state-of-the-art methods on MAE and RMSE common metrics on real public datasets.

Switch-Transformer Sentiment Analysis Model for Arabic Dialects That Utilizes a Mixture of Experts Mechanism

In recent years, models such as the transformer have demonstrated impressive capabilities in the realm of natural language processing. However, these models are known for their complexity and the substantial training they require. Furthermore, the self-attention mechanism within the transformer, designed to capture semantic relationships among words in sequences, faces challenges when dealing with short sequences. This limitation hinders its effectiveness in five-polarity Arabic sentiment analysis (SA) tasks. The switch-transformer model has surfaced as a potential substitute. Nevertheless, when employing one-task learning for their training, these models frequently face challenges in presenting exceptional performances and encounter issues when producing resilient latent feature representations, particularly in the context of small-size datasets. This challenge is particularly prominent in the case of the Arabic dialect, which is recognized as a low-resource language. In response to these constraints, this research introduces a novel method for the sentiment analysis of Arabic text. This approach leverages multi-task learning (MTL) in combination with the switch-transformer shared encoder to enhance model adaptability and refine sentence representations. By integrating a mixture of experts (MoE) technique that breaks down the problem into smaller, more manageable sub-problems, the model becomes skilled in managing extended sequences and intricate input–output relationships, thereby benefiting both five-point and three-polarity Arabic sentiment analysis tasks. The proposed model effectively identifies sentiment in Arabic dialect sentences. The empirical results underscore its exceptional performance, with accuracy rates reaching 84.02% for the HARD dataset, 67.89% for the BRAD dataset, and 83.91% for the LABR dataset, as demonstrated by the evaluations conducted on these datasets.

Bi-Fusion of Structure and Deformation at Multi-Scale for Joint Segmentation and Registration.

Medical image segmentation and registration are two fundamental and highly related tasks. However, current works focus on the mutual promotion between the two at the loss function level, ignoring the feature information generated by the encoder-decoder network during the task-specific feature mapping process and the potential inter-task feature relationship. This paper proposes a unified multi-task joint learning framework based on bi-fusion of structure and deformation at multi-scale, called BFM-Net, which simultaneously achieves the segmentation results and deformation field in a single-step estimation. BFM-Net consists of a segmentation subnetwork (SegNet), a registration subnetwork (RegNet), and the multi-task connection module (MTC). The MTC module is used to transfer the latent feature representation between segmentation and registration at multi-scale and link different tasks at the network architecture level, including the spatial attention fusion module (SAF), the multi-scale spatial attention fusion module (MSAF) and the velocity field fusion module (VFF). Extensive experiments on MR, CT and ultrasound images demonstrate the effectiveness of our approach. The MTC module can increase the Dice scores of segmentation and registration by 3.2%, 1.6%, 2.2%, and 6.2%, 4.5%, 3.0%, respectively. Compared with six state-of-the-art algorithms for segmentation and registration, BFM-Net can achieve superior performance in various modal images, fully demonstrating its effectiveness and generalization.

DopUS-Net: Quality-Aware Robotic Ultrasound Imaging Based on Doppler Signal

Medical ultrasound (US) is widely used to evaluate and stage vascular diseases, in particular for the preliminary screening program, due to the advantage of being radiation-free. However, automatic segmentation of small tubular structures (e.g., the ulnar artery) from cross-sectional US images is still challenging. To address this challenge, this paper proposes the DopUS-Net and a vessel re-identification module that leverage the Doppler effect to enhance the final segmentation result. Firstly, the DopUS-Net combines the Doppler images with B-mode images to increase the segmentation accuracy and robustness of small blood vessels. It incorporates two encoders to exploit the maximum potential of the Doppler signal and recurrent neural network modules to preserve sequential information. Input to the first encoder is a two-channel duplex image representing the combination of the grey-scale Doppler and B-mode images to ensure anatomical spatial correctness. The second encoder operates on the pure Doppler images to provide a region proposal. Secondly, benefiting from the Doppler signal, this work first introduces an online artery re-identification module to qualitatively evaluate the real-time segmentation results and automatically optimize the probe pose for enhanced Doppler images. This quality-aware module enables the closed-loop control of robotic screening to further improve the confidence and robustness of image segmentation. The experimental results demonstrate that the proposed approach with the re-identification process can significantly improve the accuracy and robustness of the segmentation results (Dice score: from <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$0.54$</tex-math> </inline-formula> to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$0.86$</tex-math> </inline-formula> ; intersection over union: from <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$0.47$</tex-math> </inline-formula> to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$0.78$</tex-math> </inline-formula> ). <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —The Doppler signal is important for the diagnosis of vascular disease, e.g., peripheral arterial disease, in clinical practices, nevertheless it is not of similar significance for state-of-the-art robotic ultrasound (US) examination systems yet. This paper explores various neural network structures to effectively extract the blood vessels from US images by incorporating the Doppler signal into the segmentation process. The final DopUS structure with two encoders extracting differentiated information from two different inputs and fusing the latent feature representations in the bottleneck layer can also inspire other tasks like multi-senor fusion. In addition, this work developed a Doppler-based tracker to assess the quality of the segmentation results in real-time. The assessment is subsequently used for a quality-aware module that enables closed-loop control of the robotic screening. Preliminary physical experiments suggest that the quality-aware robotic screening system can improve the confidence and robustness of autonomous US examination results. In the future, the Doppler signal could also be used to support clinical diagnosis. We believe the proposed quality-aware autonomous screening system is important for the development of large-scale robotic US screening programs. It will not only benefit the examination of limb arteries but also other vascular structures, e.g., carotid or aorta.

Transformer Text Classification Model for Arabic Dialects That Utilizes Inductive Transfer

In the realm of the five-category classification endeavor, there has been limited exploration of applied techniques for classifying Arabic text. These methods have primarily leaned on single-task learning, incorporating manually crafted features that lack robust sentence representations. Recently, the Transformer paradigm has emerged as a highly promising alternative. However, when these models are trained using single-task learning, they often face challenges in achieving outstanding performance and generating robust latent feature representations, especially when dealing with small datasets. This issue is particularly pronounced in the context of the Arabic dialect, which has a scarcity of available resources. Given these constraints, this study introduces an innovative approach to dissecting sentiment in Arabic text. This approach combines Inductive Transfer (INT) with the Transformer paradigm to augment the adaptability of the model and refine the representation of sentences. By employing self-attention SE-A and feed-forward sub-layers as a shared Transformer encoder for both the five-category and three-category Arabic text classification tasks, this proposed model adeptly discerns sentiment in Arabic dialect sentences. The empirical findings underscore the commendable performance of the proposed model, as demonstrated in assessments of the Hotel Arabic-Reviews Dataset, the Book Reviews Arabic Dataset, and the LARB dataset.